Who will win the Web services war?




Opinion by Brian Bakker, Computing South Africa




JUNE 23, 2004 - Technology battles have long been a feature of the IT industry. Most IT
managers will remember the LAN wars -- Ethernet .vs Token Ring -- and the bus wars --
Industry Standard Architecture .vs Micro-Channel Architecture. In both cases the eventual
winner was the least proprietary solution.

In the Web services war, the two dominant platforms are Java and .Net. Conventional
wisdom indicates that most enterprises will adopt strategies featuring both, but recent
research by Evans Data Corp. suggests otherwise. In surveying 500 developers currently
involved in Web services, Evans found that 70% believe that support for multiple operating
systems is more important than support for multiple languages.

The reasons are apparent from answers to a question about time frames for developing
Web services with wireless devices. Almost 40% of respondents are either already rolling
out Web services to such devices, or plan to do so within 12 months. A further 21.2%
expect to be extending Web services in this manner 12 to 24 months from now.

The most interesting aspect revealed by the research was the rating that developers gave
to various characteristics of Web services. The most important factor for 58% of the
respondents was support for the latest standards -- specifically those related to security
and reliability.

Other characteristics rated highly were: integration with specific Web or applications
servers (47.1%); debugging and test features (43.4%); a robust and usable GUI (43.4%);
certified support for interoperability (42.8%); and open-source support (34.9%).

On the issue of standards, Gartner Inc. analyst Daryl Plummer notes that Microsoft Corp.
has submitted approximately 40% of the APIs and specifications for .Net to the European
Computer Manufacturers Association standard body.

"However, Microsoft only submitted the lower-level API and has retained proprietary control
over higher-level elements of the platform. Nevertheless, this has successfully created the
'illusion' of open-standards support for the platform," he says.

Gartner makes the point that applications and services deployed within the .Net framework
aren't tied directly to the underlying operating system. "Although Microsoft has made no
official announcements regarding its intention to support the full .Net framework beyond
Windows, it has frequently alluded to this option," Plummer says.

Advocates of Microsoft's vision for Web services frequently mention the Common
Language Runtime element of .Net. Evans has this to say: "Through the CLR, Microsoft's
.Net supports multiple languages, but so does Java. In .Net, compiled code -- whether from
C#, VB.Net, or even Cobol -- is the same. This compiled code, called Intermediate
Language, is analogous to Java Bytecode. Equally, however, for Java, there are many
languages that code can be written in, and there are also free open-source compilers for the
Java Virtual Machine."

W ho will win the war? The answer is debatable, but, where the relative maturity of Java is a
factor in its favor, Microsoft owns the desktop, and its marketing muscle should never be
discounted.

The likelihood is that Java will continue to permeate the enterprise -- primarily at the back
end -- while Microsoft will remain a factor on the desktop, as well as in the small and
medium-size enterprise sector, where multiple platforms are less prevalent.